Reinforcing Precursors to Shift Contingency Strength Toward the Mand

You can climb a response class hierarchy by reinforcing the precursor first, then fading that precursor into a mand, all without ever putting the severe behavior on extinction. That is the whole trick. It is not a trick, really. It is a clean piece of math that a Board Certified Behavior Analyst (BCBA) can run in a supervision session and see the contingency strengths shift in the right direction before a tech ever touches the program.

This page is the worked example. We will move through it the way Matt Harrington walks through it on the recording. If you want the full CEU with the equations on screen, watch the talk.

What a response class hierarchy is in plain words#

A response class hierarchy is just a list of behaviors that all earn the same reinforcer. They show up in an order. Small stuff first. Big stuff last. The client uses whichever one works.

An ordered sequence of behaviors, and the precursor would be assent withdrawal or frowning or something like that.

Picture a kid in a classroom. A demand goes up. First, the kid pulls assent. Maybe he frowns. Maybe his arms cross. If nothing changes, he grunts. Then maybe a desk gets flipped. Then maybe a teacher gets hit. Every step in that chain is reaching for the same thing, which is escape from the demand or attention or a tangible. The small behaviors and the big behaviors are members of the same family.

The smallest behavior in the chain is the precursor. The biggest behavior is the severe behavior. They are not separate problems. They are the same problem at different volumes.

If you read this back to the tech, they will tell you they already know. They watch the kid every day. They can name the chain in order. The job of the BCBA is to write that order down and then use it.

Why precursors are the cheapest leverage point in a session#

Severe behavior is expensive. Someone gets hurt. The room clears. The data sheet stops. Recovery takes thirty minutes. You cannot iterate on a program that only runs every other Tuesday.

Precursors are cheap. They happen all day. They are easy to see and easy to respond to. A frown can be answered in two seconds. A desk flip cannot.

When you do the math, the cheap behavior is also where you have the most opportunities to deliver reinforcement. More opportunities means more data. More data means the contingency space starts to fill in. Once the precursor is the loudest behavior in the room, the severe behavior has nowhere to go.

This is the move that lets Functional Communication Training (FCT) work later. You are not skipping FCT. You are setting it up. The precursor is the bridge.

Step one: reinforce the precursor and watch the math#

In the talk, Matt sets up a scenario where severe behavior is already getting some reinforcement, the mand is getting some reinforcement, and the team is stuck. Adding more mand training is not closing the gap. So you add a third lane. You reinforce the precursor on purpose.

Yes, on purpose. This part makes people flinch. You are telling a tech, when this kid frowns and crosses his arms, respond. Give the thing. The frown earns the break. Not the desk flip.

What does the math do? Severe behavior drops to one occurrence in the session. Contingency strength for severe behavior lands somewhere around 0.2 to 0.3. That is low. That is barely a problem. Meanwhile the precursor is getting answered all over the place. Its contingency strength climbs to 0.3 or 0.4. The mand stays in the picture but it is not the headline yet.

This is the part of the run where supervisors get nervous. You are reinforcing a behavior that lives in the same family as the thing you are trying to reduce. That is the point. You are pulling the response budget out of the bottom of the chain and parking it at the top.

The tech does not need a long explanation. The tech needs a clean rule. When the precursor shows up, this is what you do. Write it down. Run it the same way every session.

Step two: put the precursor on extinction and prompt a mand#

Once the precursor is the loudest behavior in the room, you swap the rule. The precursor no longer pays out. The mand does. And you do not wait for the mand to happen on its own. You prompt it.

At no point have we put severe behavior on extinction. We're going to put the precursor on extinction.

Read that again. Severe behavior is not on extinction. You have not changed the response to the desk flip or the hit. If one happens, you still respond the way you have been. You are only changing what you do with the frown.

Here is how Matt scripts it. Demand goes up. The client frowns and crosses his arms. The tech says, looks like you need a break, let's go take a break, buddy, sound good. The client says break. The break is delivered. The mand was prompted off the precursor. The precursor itself did not earn the break this time. The mand did.

After a few rounds of this, the client stops bothering with the frown. The mand shows up first. The chain has been rewritten. The bottom of the hierarchy has a new behavior in it, and that new behavior is a word.

What the final contingency strengths should look like#

If you ran this clean, the math at the end of the session looks like this. The mand sits at a contingency strength of about 0.7 to 0.8. Severe behavior sits down at 0.1 to 0.2. The precursor is right there with the severe behavior, around the same range.

That is the same shape you would have gotten from a textbook FCT run with the severe behavior on extinction. Same finish. Different road.

The same effects of FCT without ever putting the client on extinction.

The reason that matters is fidelity. Putting severe behavior on extinction is a high cost play. The burst gets ugly. The tech has to hold the line during the burst or you make the chain stronger. If the team cannot hold the line, you have just trained the kid that escalation works. Skipping that bet, when you can, is good practice.

When this beats classic FCT with extinction on the severe behavior#

You do not always need this path. Classic FCT with extinction on the severe behavior works great when the severe behavior is mild enough to ride out, when the staff are trained tight, and when the setting can absorb a burst.

You reach for the precursor reinforcement path when one of those is not true. The severe behavior is too big to ride out safely. Staff fidelity is shaky. The setting is a public classroom with twenty other kids in it. Any of those three and you start looking for a way to move contingency strength without ever asking the team to hold extinction on a desk flip.

There is a second case. Sometimes the severe behavior is so rare that you cannot get a clean extinction run on it. You only see the desk flip every fourth session. You cannot build a strong FCT contingency against a behavior you barely have data on. The precursor happens every five minutes. Use that.

How to teach a tech to spot the precursor#

This is where the program lives or dies. If the tech cannot reliably name the precursor in the moment, the whole plan is on paper only.

Run a one page operational definition. Write three to five exact body cues. Frown. Arms crossed. Body turned away from the desk. Pencil down. Shoulders dropped. Whichever ones this client actually does. List them. No vague language. No clinical jargon.

Then watch ten minutes of video together. Pause every time a precursor shows up. Have the tech call it before you call it. Then have the tech call it live in session while you sit in the room. Then have them call it live without you in the room and check the data later.

This is also a good supervision artifact. Save the operational definition, the video review notes, and the first session of independent calls. You will want them later when a new tech rotates in or when a parent asks why the team is reinforcing a frown.

FAQ#

What is a response class hierarchy?

It is a group of behaviors that all earn the same reinforcer, sitting in an order from small to big. A frown, a grunt, a desk flip, and a hit can all be after the same break from a demand. They are members of the same response class.

Is it okay to reinforce a precursor on purpose?

Yes, inside a written plan. You are using the precursor as a stepping stone to move contingency strength toward the mand. You do not leave the precursor reinforced forever. Step one builds the lane. Step two replaces it with the mand.

How is this different from regular FCT?

Regular FCT puts the severe behavior on extinction while the mand is taught. This path never puts severe behavior on extinction. It uses the precursor as the working surface. You end up at the same place. The mand is strong, the severe behavior is weak, the contingency strengths are split.

What if the precursor is hard to identify?

Sit with the tech and the family for a week and collect video. Most clients have two or three body cues that show up before the severe behavior every time. If you really cannot find one, the response class may not be hierarchical for this client. Reach for a different equation in the talk.

When do I stop reinforcing the precursor?

The moment the mand starts showing up reliably on prompt, you stop. The precursor goes on extinction at that point. If you keep paying the precursor, you have made the precursor the new mand. That is not the goal.

Watch the talk#

This walkthrough is one piece of a longer CEU on three equations BCBAs use to predict behavior change. The full recording shows the contingency space analysis on screen, the matching law, and the percentile schedule, all with real client examples.

Watch Prediction and Probabilities with Matt Harrington on openceu.com